Textfyre presents Eric Eve's Nightfall

On Dec 17, 4:48 pm, Erik Temple <Ek.Tem....TakeThisOut@gmail.com> wrote:
> Very nice. Very slow on the pathfinding, though--the delay was so long
> I thought the game had crashed. Is this a speed issue with all
> computationally complex operations?

We're going to profile this and see if we can improve performance.

> Also note that the tab says "Table of Content" rather than "Contents";
> and "Continue Game" from the TOC brings you to the intro page, rather
> than the gameplay page. These minor issues aside, FyreVM looks
> extremely promising--thanks for sharing it!

It's still in beta stage (the UI).

You (anyone) can share your thoughts here or if you want to be more
direct, please feel free to send me your comments.

Thanks,

David C.

On Dec 17, 4:59 pm, "Eric Eve" <eric.... RemoveThis @NOSPAMhmc.ox.ac.uk> wrote:

> It's *very* slow the first time it uses pathfinding. After that it
> should it be a bit quicker.
>
> -- Eric

Ah yes, it does speed up substantially after the first time, and the
delay is not at all unreasonable. Would it be possible to pre-cache
the pathfinding during startup? I think Alabaster is doing something
like that with conversation threading.... Maybe the FyreVM crew will
be able to speed things up in any case, but the delay is really
substantial--I closed my browser window the first time, thinking the
game had tanked...

--Erik

On Dec 17, 4:18 pm, ChicagoDave <david.cornel....RemoveThis@gmail.com> wrote:
> From the call for samples, Eric Eve chimed in and offered his latest
> competition game as a guinea pig.
>
> We've still got some technical issues to resolve, but the game is
> mostly playable at this point. There is a map for Nightfall at <http://
> ifdb.tads.org/viewgame?id=3a4kqs374jqu2li6> if you need one.
>
> Without further ado, here is an Interactive Fiction presented in
> Textfyre's FyreVM using Channel IO published in Microsoft Silverlight
> and of course thanks to Eric Eve for the gracious use of his second
> place IFComp 2008 game.
>
> Note: Silverlight works on Windows XP, Vista, and Intel Mac's using
> any of the common browsers.
>
> Nightfall: <http://www.textfyre.com/nightfall>

If you're on a Mac and get an error saying something about a wrong
beta version, download Silverlight directly:

http://silverlight.net/GetStarted/

And then go back to the nightfall page.

David C.

This is pretty, but I still want to grab the protagonist and just
shake him

UI comments:

* When the first prompt appeared, I was busy clicking pages past and
then suddenly that stopped working. It only took me a fraction of a
second to go "duh, we're at the part with a parser," but some kind of
helpful reminder to that effect maybe should pop up when someone
clicks the lower-right "page turn" area when the parser is expecting
input for the first time (or maybe not, if the plan is to explain this
stuff up front in the finished dealy).

* In my browser (a current Firefox install) there were some very ugly
kerning problems, almost as if the program were kerning for the wrong
typeface. Some letters were head-butting each other; others were
clearly having a tiff and standing with an icy distance between them.

* I played with returning to the "Table of Content" from the game, and
then from there "Return to Game," but this didn't bring me back to the
game proper. Instead, it brought me back to the next "page," and from
there I had to manually "flip" to the actual game screen.

Other than that, just purty!

On Dec 17, 5:01 pm, ChicagoDave <david.cornel....DeleteThis@gmail.com> wrote:
> On Dec 17, 4:48 pm, Erik Temple <Ek.Tem....DeleteThis@gmail.com> wrote:
>
> > Very nice. Very slow on the pathfinding, though--the delay was so long
> > I thought the game had crashed. Is this a speed issue with all
> > computationally complex operations?
>
> We're going to profile this and see if we can improve performance.
>
> > Also note that the tab says "Table of Content" rather than "Contents";
> > and "Continue Game" from the TOC brings you to the intro page, rather
> > than the gameplay page. These minor issues aside, FyreVM looks
> > extremely promising--thanks for sharing it!
>
> It's still in beta stage (the UI).
>
> You (anyone) can share your thoughts here or if you want to be more
> direct, please feel free to send me your comments.

A couple of other things I noticed:

-- if you try to go through a locked door, you get a custom message
from the game, then a y-umlaut (’), and then the default "(first
trying to open the street door)" sort of message. I'm not sure what's
going on there, but I don't think anything like that happened in the
comp version of the game.

-- this may be just me, but having the uncurved text/images against
the picture of a strongly-curved book background is a little
disconcerting. I know you probably don't want to bend the lettering to
follow the shape of the "pages" behind, but I thought this a mildly
weird effect.

On the other hand, the images do give a sense of the city as a really
large place -- larger even than I had formed in my head while playing
the first couple of times. That's an interesting effect for the game.

> It is not just you: as soon as I noticed this it continued to bother me
> as long as I played. I can't suggest a proper fix for this, but it
> might be less disconcerting if the pages were pushed down a little
> flatter. (I suppose this would be the place to say that I'm a little
> dubious about the book metaphor anyway. I'm not entirely sure what it
> gains the player.)

After the initial couple of "pages," it doesn't seem to do anything at
all anymore (maybe there's a later part in the game where the page
"turns" again to mark the passing of a notable event? I only played
for a little while ...)

If the pages never mean anything past the intro, it might be best to,
at the last "turning," simply provide a cute little zoom-in so that
the "page" graphic gets out of the player's way. Dunno.

Or in other words: ditto on the dubious. Pretty, though.

On 2008-12-17 20:17:58 -0500, Emily Short <emshort.DeleteThis@mindspring.com> said:

> On Dec 17, 5:01 pm, ChicagoDave <david.cornel....DeleteThis@gmail.com> wrote:
>> [...]
>>
>> You (anyone) can share your thoughts here or if you want to be more
>> direct, please feel free to send me your comments.
>
> A couple of other things I noticed:
> [...]
> -- this may be just me, but having the uncurved text/images against
> the picture of a strongly-curved book background is a little
> disconcerting. I know you probably don't want to bend the lettering to
> follow the shape of the "pages" behind, but I thought this a mildly
> weird effect.

It is not just you: as soon as I noticed this it continued to bother me
as long as I played. I can't suggest a proper fix for this, but it
might be less disconcerting if the pages were pushed down a little
flatter. (I suppose this would be the place to say that I'm a little
dubious about the book metaphor anyway. I'm not entirely sure what it
gains the player.)

> [...]

Best,
James

"Erik Temple" <Ek.Temple.RemoveThis@gmail.com> wrote in message
news:fcfda271-02f8-495a-bb22-2e1db65afdd0@n33g2000pri.googlegroups.com...
On Dec 17, 4:18 pm, ChicagoDave <david.cornel....RemoveThis@gmail.com> wrote:

> Very nice. Very slow on the pathfinding, though--the delay was so
> long
> I thought the game had crashed. Is this a speed issue with all
> computationally complex operations?

It's *very* slow the first time it uses pathfinding. After that it
should it be a bit quicker.

-- Eric

ChicagoDave <david.cornelson RemoveThis @gmail.com> wrote:
> Without further ado, here is an Interactive Fiction presented in
> Textfyre's FyreVM using Channel IO published in Microsoft Silverlight
> and of course thanks to Eric Eve for the gracious use of his second
> place IFComp 2008 game.

My first impression: That's a *huge* font. I feel like I'm trying to
read a book while peering through a narrow cardboard tube.

- Damien

> -- if you try to go through a locked door, you get a custom
> message
> from the game, then a y-umlaut (’), and then the default "(first
> trying to open the street door)" sort of message. I'm not sure
> what's
> going on there, but I don't think anything like that happened in
> the
> comp version of the game.

Actually, I get that a lot in my WIP - got it ever since I added the
ASCII automap. It's surely the extension, and I don't know why it's
happening. At any rate, it's not a FyreVM issue.

BTW, this has happened in Glulx. Maybe it doesnt' happen in a Z-
machine?

> I think the
> y-umlaut is caused by the print char(255) that Text Capture uses to
> mark the end of a captured string; it's being displayed on the
> screen instead of added to the text buffer. The stuff that's being
> printed before the "(first trying to open the street door)" is
> probably text that should have been captured but wasn't.

The y-umlaut *is* the stuff being printed before the text, and since
there's no time at all in my WIP where I display such a character (or
even in your extension - I browsed through), I don't know why it
appears at all.

> I'm not sure why it would be happening in your WIP if you're WIP
> isn't using the FyreVM, unless the implementation of your ASCII
> automap is incompatible with Text Capture. Does the ASCII automap do
> anything with output text capturing?

Very simply, I fire a set of rules when the player moves or opens/
closes a door, and those rules set the focus to a separate window
(using Flexible Windows by Jon Ingold), which is a text grid, and then
I write a bunch of stuff on the text grid.

I've noticed the problem persistently, but *not at all times*. It
seems to appear at certain points, or after I've visited a certain
location. Once it's there, I can replicate it easily, but I haven't
yet bothered to find the conditions which make this happen.

> I noticed this too. This applies to all implicit actions, not just
> going through doors, and, as you say, didn't occur in the comp
> version of the game and doesn't occur in an ordinary 'terp. I think
> at least part of the problem is that my Text Capture extension,
> which Implicit Actions uses, doesn't work properly with the FyreVM.

I've had this bug when trying to blend Text Capture and Flexible
Windows. The umlaut character is one printed by Text Capture to mark
the end of the captured text: the problem is [something like] when a
change of window is issed, the capturing process stops silently, so
when the end capture is reached, it prints the character to the screen
rather to the buffer. I managed to get round it by ending streams
before changing windows, but I never solved it.

For the pathfinding: could you do a "go to <X> silently" at start-up?
Presumably the problem is that the particular path-finding routine
needs to be fired, so all you need to do is ensure that when you call
it during startup it's going through the same line of I7 code?

jon

"Erik Temple" <Ek.Temple DeleteThis @gmail.com> wrote in message
news:f7dbf369-a560-45c3-90af-c5324ae4bb5b@q26g2000prq.googlegroups.com...
On Dec 17, 4:59 pm, "Eric Eve" <eric.... DeleteThis @NOSPAMhmc.ox.ac.uk> wrote:

> Ah yes, it does speed up substantially after the first time, and
> the
> delay is not at all unreasonable. Would it be possible to
> pre-cache
> the pathfinding during startup?

Unfortunately I can't see how. That is, I've tried issuing a
pathfinding command at startup for just this purpose, but the only
result is that the delay occurs twice over, once at startup and once
the first time the player uses a GO TO/CONTINUE command that uses
pathfinding.

-- Eric